[PATCH] D59095: [llvm-readelf]Don't lose negative-ness of negative addends for no symbol relocations

[James Henderson via Phabricator via llvm-commits]

jhenderson marked 2 inline comments as done.
jhenderson added a comment.

Thanks all. I am already planning a number of follow-up changes to fix the issues re. consistent formatting and the LLONG_MIN comment. This particular issue however is something I needed more urgently, so wanted to fix first.



================
Comment at: tools/llvm-readobj/ELFDumper.cpp:2735
+        Addend = " - ";
+        RelAddend = std::abs(RelAddend);
+      } else
----------------
MaskRay wrote:
> If `r_addend = LLONG_MIN`, `abs(r_addend)` is ub...
> 
> "The behavior is undefined if the result cannot be represented by the return type."
Yes, it is. But it's already a problem, so I'd like to fix this separately, if that's okay?


================
Comment at: tools/llvm-readobj/ELFDumper.cpp:3401
+    }
+    Addend += to_string(format_hex_no_prefix(RelAddend, 1));
   }
----------------
Higuoxing wrote:
> Can we use `to_hexString(RelAddend, false)` here? I think it's better to have consistent code style.
Would you mind if I do this in a later patch? I'm looking to try to share the code for `printRelocation` and `printDynamicRelocation`, since they're almost identical anyway, and this would be one of the changes I'm planning on then.


Repository:
  rL LLVM

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D59095/new/

https://reviews.llvm.org/D59095